Real data is messy. If it's because of changes in standards, human error, or just because the real data points are messy, when taking real data from insitutions that were usually not keeping out around for the purposes of data science, we tend to encounter inconsistent formatting, missing values, duplicates, inconsistent typing, and other issues.
dftest (inspired by pytest) is a project aims to give data scientists tools to detect problematic data which may lead to unexpected results, and loctae these rows and columns which may need to be removed or require additional cleaning.
Full documentation at README.ipynb. Some of the outputs in the notebooks are partial or incorrect - specifically, graphs don't display correctly and the save action shows errors. This does not reflect actual behaviour and will be fixed in the coming days.